Will patents pillage open source?

Open-source software is indeed vulnerable to claims of patent infringement. But attorney Steven Frank says that doesn't mean the legal floodgates are about to open.

4 min read
Is open-source software just one patent away from annihilation? Many people think so and point to the ominous convergence of two trends: the widening availability of software patents and the willingness of intellectual-property owners to challenge distributors of open-source software. If someone can obtain a patent to key features of the Linux kernel, couldn't that person bring the entire open-source community to its knees by threatening industrywide legal action? At that point, shouldn't CIOs flee open source for the safety of Microsoft and Sun Microsystems?

Not so fast.

To be sure, open-source software is as vulnerable to claims of intellectual-property infringement as any other category of software; the billion-dollar lawsuit brought by SCO Group--which inherited intellectual-property relating to the Unix operating system--against IBM demonstrates that. Indeed, SCO's intellectual-property claims involve trade secrets, which are usually far less potent weapons than patents. Its lawsuit is unlikely to have much impact on Linux, and the threat posed by patents is, if anything, more remote.

As a business strategy, squeezing the open-source community for license royalties leaves a lot to be desired. Suddenly unleashing IP claims on a large installed base isn't exactly conductive to attracting customers and business partners. Many analysts, for example, view SCO's lawsuit as the last desperate gasp of a marketplace loser; in its quarter ended Jan. 31, SCO posted a net loss of $724,000 on revenue of $18 million.

Such a strategy seems particularly misplaced in the Linux arena.

Open-source adopters relinquish a measure of proprietary independence (and traditional intellectual-property royalties) in order to benefit, without charge, from the innovative efforts of others. The imposition of a monetary tax strikes at the heart of the open-source bargain; cheerful cooperation on the part of Linux users is, to put it mildly, unlikely.

Notice whom SCO chose to sue--not a large Linux user, not a product vendor such as Red Hat or SuSE, but Big Blue. Intellectual-property litigation is very expensive. As a practical matter, therefore, the ability to enforce intellectual-property requires a relatively small number of infringers that dominate the market, and which have resources justifying the legal effort. But open-source users are a huge and widely diverse lot who can always resort to home-grown alternatives. Besides, how much in the way of legal damages could be recovered from any particular user?

The SCO case is somewhat unusual in its focus on trade secrets. Like most trade-secret actions, this one arose from a corporate relationship gone awry, with SCO alleging misuse of information disclosed in confidence. The relative rarity of such situations in open-source contexts (which are, after all, "open") suggests that future disputes will probably involve patents or copyright--and most likely patents.

The reason is twofold. Copyright infringement requires actual copying, while an independent developer can be liable for patent infringement. In addition, the scope of copyright protection for software has narrowed considerably over the last decade.

Software patents sometimes cause legitimate controversy not because computer programs somehow differ from other patentable technologies, but because patents on software are relatively new. Undergirding every patent is faith that the U.S. Patent and Trademark Office has performed a competent literature search, and awarded patent protection only to subject matter that's "inventive"--i.e., new and different from prior work. That faith has been sorely tested in the software arena.

The Patent Office focuses its literature search largely on prior patents, because patents are what is most familiar and easily accessible. But because comparably little prior software work has been patented, the search may be incomplete, so patents that should never see the light of day sometimes do.

The open-source community has at its disposal a wealth of information on such technology, putting it in a good position to invalidate misbegotten software patents.

For that very reason, however, those patents remain vulnerable to attack; the defendant in an infringement lawsuit can always challenge the validity of a patent based on technology that had previously been published or sold-- particularly technology unknown to the Patent Office when it granted the patent. The open-source community has at its disposal a wealth of information on such technology, putting it in a good position to invalidate misbegotten software patents.

Moreover, a community of developers could collaborate to quickly develop a noninfringing work-around. Such threats, added to the high costs of litigation, questionable returns and the prospect of widespread rejection of open wallets for open source, will likely engender hesitation among most patent owners.

Recent experience involving industry standards provides a useful parallel. For years, participants in standards-setting processes have been obtaining patents on technology incorporated into standards, and routinely charging royalties to all users. The World Wide Web Consortium decided to make itself an exception. Last fall, it all but formally adopted a policy of requiring contributors to make IP available on a royalty-free basis.

Almost immediately, Epicentric, a unit of Vignette that earlier had claimed patent coverage relevant to a W3C standard, backed off its threat to charge royalties under the patent. Similarly, IBM last year said it will not seek royalties on patented technology involving a different Web standard (which pertains to e-commerce).

With more and more innovation companies retreating even in circumstances traditionally justifying license royalties, the prospect that patents could doom open source grows ever more unlikely.